Antral gastrin- and somatostatin-producing cells and intraluminal peptide secretion in normal subjects and duodenal ulcer patients with and without vagotomy

Regulatory Peptides, 14 (1986) 133-143

133

Elsevier RPT 00469

Antral gastrin- and somatostatin-producing cells and intraluminal peptide secretion in normal subjects and duodenal ulcer patients with and without vagotomy O. Gutierrez*, E. Ren6, J.P. Accary, T. Lehy, J.P. Laigneau, J.A. Chayvialle ~ and S. Bonfils INSERM U.IO, Unit~ de Recherches de GastroentOrologie, Hrpital B&hat, 170 Boulevard Ney, 75877 Paris Cedex 18, and alNSERM U.45. Hrpital Edouard Herriot, 69374 Lyon Cedex 2, France

(Received 16 August 1985;revisedmanuscript received29 November 1985;acceptedfor publication 30 January 1986)

Summary The behaviour of gastrin (G) cells and somatostatin (D) cells in endoscopic antral biopsies and that of intraluminal gastrin (ILG) and somatostatin (ILS) release in the gastric juice were investigated in three groups of patients: control subjects, duodenal ulcer (DU) patients and DU patients treated by a superselective vagotomy (SSV). G and D cell densities were correlated in the three groups of subjects. The G/D cell ratio was significantly increased in SSV patients (P < 0.001) as compared to control and D U patients. No correlation was found between gastrin or somatostatin cell densities and basal intraluminal levels of the two peptides. ILG output was significantly higher in DU patients than in control or SSV patients (P < 0.001). ILS output was also higher in D U patients than in controls (P < 0.001) and in SSV patients (P < 0.05). It was also significantly augmented in SSV (P < 0.001) as compared to control patients. ILG and ILS concentrations were only correlated in controls. Within each of the three groups of subjects, ILG and ILS release varied in function of the gastric juice pH. Our results emphasize the necessity to consider the intragastric pH as well as the physiological or pathological state to study intraluminal peptides in man. intraluminal secretion; gastrin; somatostatin; endocrine cells; gastric acid secretion

*To whom all correspondenceshould be sent.

134 Introduction

Gastrin and somatostatin are two peptides, which have been shown to play a role in the physiological regulation of gastric acid secretion [1-3]. In the gastric mucosa, these peptides are produced in the gastrin (G) cells and in the somatostatin (D) cells, respectively. They are released into the portal vein [4,5] and into the gastric lumen. It has been shown that release of intraluminal gastrin (ILG) and intraluminal somatostatin (ILS) are both affected by the antral pH [2]. Until now, in man, few reports have dealt with the presence of ILG and ILS in the gastric juice [2,6-8]. While several studies were conducted on antral gastrin cells of the human stomach it is only recently that the G/D cell ratio has been studied and found to be dependent on gastric secretion [9]. In this work, we examined the G and D cell densities in antral biopsies as well as the ILG and ILS levels in gastric juice of normochlorhydric subjects, duodenal ulcer patients (DU) and DU patients treated by a superselective vagotomy (SSV). The possible relationships between intraluminal levels of the two peptides and either the G and D cells, the circulating gastrin and somatostatin levels, or the basal acid secretion were investigated.

Material and Methods

Patients Three groups of patients were studied: (1) 15 normal control subjects (C) without gastrointestinal pathology; (2) 19 patients with previously endoscopicaUy diagnosed DU, who were at the end of a 6-week treatment (cimetidine 1 g/24 h or ranitidine 300 mg/24 h); at that time, 10 of them were healed while 9 had a persistent endoscopic DU; and (3) 13 uncomplicated D U patients, operated on from 1 to 30 months after SSV (mean 4.1 months). The SSV was performed without drainage procedure. Among them four patients were studied before and after SSV. The quality of this operation was checked by insulin-pentagastrin test. The mean age of control patients (8 males and 7 females) was 42.4 years (range 17-79.5 years), that of DU patients (16 males and 3 females) was 40.5 years (range 18-81 years) and that of SSV patients (11 males and 2 females) was 41.4 years (range 18-80 years). All DU patients stopped anti-H2 medication 36-48 h before endoscopy. No other digestive pathology was present. All subjects were consenting and the study was approved by the Human Studies Committee, H6pital Bichat, Paris. Gastric juice collection After a 16 h fasting period a nasogastric tube was inserted and its position controlled by water test. Gastric juice was collected on ice for the next hour (4 x 15min periods) by gentle suction, pH was measured for each sample. Basal acid output (BAO) was determined. Pepstatin (10/~g/ml) was added to prevent peptic digestion. The samples were divided into two portions, one of which was adjusted to pH 7-8 with 1 M NaOH (for IRG assay), the other to pH 1-2 with 1 M HC1 (for IRS assay).

135 Then both portions were boiled 5 min to avoid proteolytic damage of the labeled antigen and antibody during radioimmunoassay. The samples were rapidly frozen, stored at -20"C and lyophilized prior to analysis for ILG and ILS. Serum samples were collected before gastric aspiration for determination of circulating levels of gastrin and somatostatin.

Radioimmunoassay of gastrin and somatostatin (1) Gastrin. We measured gastrin levels both in plasma and gastric juice, using a previously described method [10]: antiserum against human synthetic gastrin I obtained in the rabbit was used at a 1/100 000 final dilution, a25I-labeled human gastrin I (1:1-17) as tracer, and the same as standard. Separation of bound from free hormone was performed using the uncoated charcoal method after 48 h incubation. The sensitivity of the technique is 2 pg gastrin per reaction mixture. All concentrations obtained by radioimmunoassay are expressed in units equivalent to the radioactivity of heptadecapeptide human gastrin I. (2) Somatostatin. The evaluation of somatostatin in gastric juice was done using a method described by Rorstadt et al. [11], the components of which are: an antiserum raised in a sheep (a generous gift of P. Brazeau, Montreal) and used at a 1/35 000 final dilution; somatostatin-14 as standard and 125I-Tyr-1 somatostatin as tracer (both from Bachem, CA, U.S.A.). After 24 h incubation, separation of the bound from the free part was done by 90% alcohol. The sensitivity of this method was 10 pg by assay. The evaluation of plasma somatostatin level was done using previously published method [12].

Recovery and gel filtration experiments To ascertain that the rate of degradation of the two hormones was the same in the different groups of patients, we performed recovery experiments. After adding either standard human gastrin I, or synthetic somatostatin-14, we allowed the samples to boil during different times (1, 5, 10 min). The levels of IRG and IRS were then measured and the recovery ratios estimated. In the three groups, the recovery ratios were between 75 and 120% for IRG and between 60 and 74% for IRS. For each group, we also did gel filtration of gastric juice, and measured IRG and IRS in the different fractions of the chromatographies. As described elsewhere [10,13], gel filtration was performed on Sephadex G50F; Na 125I, bovine serum albumin, 125I-gastrin and ~2SI-somatostatin were used as markers. In the three groups of subjects, gel filtration profiles of immunoreactive gastrin showed a large peak of G-17, a smaller of G-34 and a large peak of big-big gastrin, at the albumin level; gel filtration profiles of immunoreactive somatostatin showed the presence of both somatostatin-14 and somatostatin-28 in the gastric juice samples.

Tissue sampling and immunohistology Immediately after gastric juice collection, the nasogastric tube was removed and a gastroscopy (Olympus GIQ) was performed. Four to five forceps endoscopic biopsies were removed from the greater curvature in the antral part, 2-3 cm above the pylorus. They were stretched on cardboard and fixed in freshly prepared Bouin's

136 solution for 24 h, dehydrated, then embedded in paraffin. 4/~m thick well-oriented tissue sections, i.e. perpendicular to the surface of the mucosa, were obtained. Gastrin (G) and somatostatin (D) cells were identified on serial sections by immunoperoxidase reaction using Sternberger's peroxidase-antiperoxidase (PAP) procedure, Tissues were incubated for 48 h at 40C with either diluted (1:1000) rabbit antigastrin serum raised against synthetic human gastrin (1:2-17) or diluted (1:1000) rabbit somatostatin antiserum raised against synthetic cyclic somatostatin. The PAP complex (Medac, Hamburg, F.R.G.) was diluted 1:100. Peroxidase activity was revealed by diaminobenzidine. Specificities of gastrin and somatostatin antisera have been described elsewhere [14,15]. Nuclei were counterstained by hematoxylin. Control staining for the specificity of the immunocytochemical reaction involved: (a) omission of the first antiserum; (b) incubation with preabsorbed antigastrin or antisomatostatin serum with an excess of antigen in place of the first layer antiserum. No positive cells were observed in the antral mucosa of control sections. Quantification of endocrine cells was performed in two ways. First, all nucleated G and D cells recognized on each biopsy section were counted under oil immersion ( x 630). The ratio of total number of G cells to total number of D cells per section was calculated. A mean for each patient was obtained by averaging individual section ratios. Second, in the well-oriented areas where the mucosa appeared in its full thickness, the densities of G and D cells per surface area (660 #m 2) were estimated under a 252 x magnification, using a calibrated ocular grid, 165/~m width [15]. Thus at least 5 different counts were performed on each biopsy section and for each patient the mean number of each cell type per area was obtained by averaging all individual counts. The ratio of G cell to D cell densities was also calculated. Results are expressed as mean ± 1 S.E.M. Statistical analyses for independent populations were performed using either the Student's t-test or the Mann-Whitney U-test whenever relevant. The paired data t-test was also employed for studying evolution of parameters within the same population. Correlations between different parameters were assessed using linear regression analyses. The level of significance was set at P < 0.05.

Results

No differences were noted between the two subgroups of DU patients with respect to intraluminal peptides as well as antral endocrine cell densities. So their data were presented as a whole in the group called DU.

Histological results The G/D cell densities were within the same range in C and DU patients (Table I). In SSV patients, G cell density increased and D cell density tended to decrease as compared to C and DU patients, but differences were not significant. In this group, however, there was a significant increase (P < 0.001) of the G/D cell ratio, regardless of the means of calculation (Table I). In the 4 patients studied before and after SSV, a clear enhancement of the G / D cell ratio was also noted after SSV (P < 0.02, Fig.

137 TABLE I G and D cells in the antrum o f the 3 groups o f patients a

(12)

Controls (14)

D U (19)

ssv

9.4 5.3 1.83 1.94

9.7 5.8 1.82 1.90

12.9 4.3 3.13 3.26

G cell density per unit area s D cell density per unit area b G/D cell densities G/D cell: total No. of cells per section

+ 44+

1.4 0.7 0.09 0.I1

4+ 44-

1.2 0.7 0,13 0,I0

+ 4+ 4-

2.1 0.7 0.30* 0.28*

No o f subjects, in whom antral biopsies were removed, are indicated between parentheses. " Biopsies removed from the greater curvature, 2-3 cm above the pylorus. b Unit surface area: 660 #m 2. * P < 0.001 vs. C and DU. G / D _ cell ratio

Intralominal Gastrin

Intraluminal Somatostatin

pgx 15mn (xtO3)

pg x 15rnn

,x,%

=k

12

8

8

before

A

after

before

• p< 0.02

B

before

after

after

"X'~ p< 0.01

Fig. 1. Evolution o f some parameters in the four duodenal ulcer patients examined before and after superselective vagotomy. (A) G/D cell ratio (total No. of cells per section). (B) Intraluminal release of peptides. 25

r:0.924 D<0.O01

"E= ,,o 09 ¢J L9

15 ~ 20.

0

10. 5-

4

25- r=0.914

r=0,67 p
25

20

20"

15

15-

10

10-

5

7

5-

/

y= 2 . 7 3 × ÷ 1 1 5

I,

No. D cells/660 ~m 2

C ftz /

D U (tg)

S S V (12)

Fig. 2. Correlations between gastrin (G) cell density and somatostatin (D) cell density within the 3 groups of patients. C: controls, DU: duodenal ulcer, SSV: duodenal ulcer with superselective vagotomy.

138 1A). A significant positive correlation (P < 0.01 to P < 0.001) between the density of G cells and that of D cells was found in the 3 groups of patients (Fig. 2).

Determinations in gastric juice (1) Gastric acid secretion. BAO was significantly higher in DU patients (7.95 ± 1.90 mmol/h) than in C (2.76 4- 0.53 mmol/h, P < 0.025) and SSV patients (0.64 4- 0.24 mmol/h, P < 0.01). The BAO was also lower (P < 0.005) in SSV patients than in controls. (2) Intraluminal peptides. Both concentration and output of ILG in the gastric juice were enhanced in the DU patients as compared to C and SSV patients. Only the difference in ILG output value observed between DU and C or SSV patients was significant (P < 0.001) (Fig. 3). Concentration and output of ILS in the gastric juice were significantly increased in both DU and SSV patients (P < 0.01 and P < 0.001, respectively) as compared to controls (Fig. 4). Furthermore, the augmentation in ILS output was more marked in DU patients than in SSV patients (P < 0.05). In the four patients studied before and after SSV, a significant decrease in ILG output was noted after operation (P < 0.01); a decrease in ILS output, although not significant, was also observed (Fig. 1B). The ratio of ILG to ILS concentration was 0.31 for controls, 0.09 for DU or SSV patients. In each gastric juice sample, variations of ILG concentrations in function of the pH were investigated (Table II). Interestingly, while mean concentrations were similar in both groups (Fig. 3), these variations were opposite in C and SSV patients: that is when the pH was greater than 3, ILG values were elevated in the C and low in the SSV patients and the contrary was observed when the pH was less than 3. The difference between the values observed at pH either superior or inferior to 3 was significant within each of these two groups (P < 0.025 in controls and P < 0.005 in SSV patients) while no significant difference was noted within the DU group. As INTRALUMINAL

GASTRIN

Output

Concentration

pg/ml

pg/l§ mn (x1021

150 -

45-

100-

30-

50-

15

C [60]

OU [72]

SSV [52]

C [6(9]

DU [72]

$$V [52]

* e p
Fig. 3. Intraluminal gastrin values in the 3 groups of patients. C: controls, DU: duodenal ulcer, SSV: duodenal ulcer with superselectivevagotomy.

139 INTRALUMINAL

Concentration

~

SOMATOSTATIN

Output

1.500]

*

60 45.

1.000

30 5

0

0

C

~

DU

[59]

[651

15-

C

SSV

[59]

[45]

* p < O O t vs C

**p
DU

[66]

,p
SSV

[451

vs SSV

Fig. 4. Intraluminal somatostatin values in the 3 groups of patients. C: controls, DU: duodenal ulcer, SSV: duodenal ulcer with superselectivevagotomy. regards I L G o u t p u t s , they varied in the same way as c o n c e n t r a t i o n s : there was a significant difference in the values observed at p H superior or inferior to 3 within the 3 g r o u p s (C group: P < 0.01; D U group: P < 0.05; SSV group: P < 0.001). N o significant difference was observed in ILS o u t p u t or c o n c e n t r a t i o n in f u n c t i o n o f the p H within a n y g r o u p (Table II).

Plasma peptide determinations P l a s m a gastrin values were o f the same order in the D U a n d SSV groups (83.6 + 6.4 pg/ml a n d 78.5 4- 8.1 pg/ml, respectively) a n d m o r e elevated in the C g r o u p (140.9 + 22.9 pg/ml, P < 0.02 c o m p a r e d to D U a n d P = 0.05 c o m p a r e d to SSV patients). TABLE I1 Variations of intraluminal gastrin and somatostatin concentrations in function of the pH of the gastric juice" ILG concentration (pg/ml)

ILS concentration (pg/ml)

C

DU

SSV

C

DU

SSV

pH < 3

61 4- 15" (28)

156 4- 30 (58)

323 4- 150"* (20)

176 + 80 (28)

1707 4- 386 (54)

859 4- 383 (17)

pH ~> 3

127 4- 22 (32)

64 + 13 (14)

51 4- 11 (32)

440 + 272 (30)

325 + 186 (12)

1169 + 310 (28)

a Gastric juice samples were collected every 15 min. Numbers of samples are indicated in parentheses. * P < 0.025 when compared to C values at pH I> 3. ** P < 0.005 when compared to SSV values at pH I> 3.

140

Plasma somatostatin values did not vary in the 3 groups (20 ± 3.7 pg/ml, 21.9 q- 6.7 pg/ml, 19.3 + 2.8 pg/ml in C, DU and SSV patients, respectively). Relationships between peptide levels, endocrine cell densities or acid secretion Peptide levels. In the controls, relationships were found between ILG and ILS concentrations (r = 0.76, P < 0.01), between ILG concentration and plasma gastrin level (r = 0.67, P < 0.01) and ILS concentration and plasma somatostatin level (r = 0.54, P < 0.05). No such correlations were found in the other two groups. Cell densities. No correlation existed between G cell density and intraluminal or basal plasma gastrin, between D cell density and intraluminal or basal plasma somatostatin, finally between G or D cell densities and acid secretion in any group. Acid secretion. Only in the SSV patients was a relationship between I L G concentration and H ÷ concentration noted, being calculated either on each 15 min sample or on the mean of the secretory tests (P < 0.02). No correlation was noted between basal plasma gastrin or somatostatin levels and gastric acid secretion.

Discussion

This study provides data concerning the intraluminal gastric release of two regulatory peptides, gastrin and somatostatin, and the relative proportion of their cells of origin, in the stomach of three groups of patients, in a basal state. We chose to take biopsies from the prepyloric region in the greater curvature since, in man, this area was found to be the richest in gastrin cells, with gastritis generally less developed [16-18]. The choice to study DU patients treated by anti-H2 blockers is based on previous works reporting that this treatment did not alter G cell densities [9] nor D cell densities [19]. Although G cell density tended to increase in our SSV patients, the difference observed as compared to C and DU patients was not significant. Real hyperplasia of gastrin cells has been found in the rat after truncal vagotomy [20] and higher G cell densities have been observed in dog [21] and man [9] after SSV. It has been previously shown that in the rat antral mucosa, the G/D cell ratio is very stable throughout adult life, with little variation between the different areas of antrum as between individuals [22]. In our study, the G/D cell ratio appeared homogeneous between individuals within a group and similar in controls and DU patients. It was of the same order as that which can be calculated in comparable groups of patients from a recent article [19]. We found this ratio to be significantly enhanced in SSV patients, confirming another report [9], although the magnitude of the increase was less important. This increase in G/D cell ratio indicates a modification in the behaviour of these two endocrine cell populations which seem to have regulatory interaction on their secretory activity. Furthermore, a significant correlation was found between antral G and D cell densities in C, DU but also in SSV patients (r = 0.91). This last fact is at variance with the results of Arnold et al. [9]. We did not find any difference in basal serum gastrin levels between DU and SSV patients. The latter fact is in agreement with previous studies where no elevation of basal gastrinemia in patients after truncal vagotomy [23] as well as SSV [23,24] was

141 reported. In man, contradictory results have been reported concerning the relationships existing between fasting serum gastrin and basal acid output [9,25] or G cell density [25,26]. Our results confirm those of Arnold et al. [9] in that augmented acid production in DU patients seems not to be due to any change in the relationship between G and D cells. The latter is the same as that found in control patients. To our knowledge, it was the first time that intraluminal gastrin and somatostatin releases were concomitantly studied in basal state in control subjects as well as in non-operated and SSV-operated DU patients. One of us had previously demonstrated that basal ILS release was higher in SSV patients than in normal control subjects [8]. In the present work, this fact was confirmed but it was also shown that ILS release was still higher in DU than in SSV patients. Moreover, ILG release was also enhanced in DU patients as compared to controls and SSV patients; in those last two groups intraluminal levels were comparable. The decrease in ILG and ILS output found in SSV patients as compared to DU patients was confirmed in the four DU patients examined before and after SSV. It should be noted that the large individual variations we observed for the two peptides have already been reported for ILG in the cat [27]. The attempt to find a simple correlation between intramucosal endocrine cells, and either intraluminally secreted peptides or plasma levels failed since the only correlations concern ILG and ILS secretions in the C group. However, we could point out several facts which might have physiological or pathological meaning. First, gastric juice pH has an effect on intraluminal peptides. The variation in ILG secretion as a function of the pH in normal subjects is the opposite of the one we observe in SSV and DU patients. Our data emphasize the necessity to consider the intragastric pH and the physiological or pathological state to compare intraluminal peptide values among subjects. Second, there is a contrast between the high ILS output in DU patients and the high acid production, since the intragastrically administered exogenous somatostatin inhibits gastric acid secretion [6]. Thus, the elevated acid production in DU patients cannot be due to a decrease in ILS secretion. Nevertheless we cannot rule out a defect in somatostatin secretion modulation by the intraluminal pH. It has been suggested that the inhibition of gastrin release and of acid secretion by a low intragastric pH is defective in patients with DU [28]. In this respect, our findings that ILG and ILS concentrations are correlated only in the C group and not in the DU and SSV patients may indicate another defect in peptide regulation in patients with DU. The high secretion o f l L S in SSV patients as compared to controls cannot be explained by a low antral pH. Indeed ILS release is known to be stimulated by an acidic pH [2]; however vagotomy induces a marked reduction of gastric acid secretion [29]. Thus ILS elevation in SSV may be explained by other mechanisms such as a change in the secretion of the D cells present in the denervated fundic area, or alternatively, an enhancement of the sympathetic nervous system which could follow gastric vagal denervation [30].

142

Acknowledgements The a u t h o r s acknowledge Professeur M. M i g n o n for giving them the o p p o r t u n i t y of studying some patients. They t h a n k Mrs. L. Gr6s, D. Labeille, C. Poitevin a n d Mr. R. Teyssedou a n d G. Peranzi for their excellent technical assistance a n d Mrs. M. S a u v a d e t for her help in p r e p a r i n g the m a n u s c r i p t . This research was s u p p o r t e d by the I n s t i t u t N a t i o n a l de la Sant6 et de la Recherche M6dicale, I N S E R M , F r a n c e .

References 1 Bloom, S.R., Mortimer, C.H. and Thorner, M.O., Inhibition of gastrin and gastric acid secretion by growth-hormone release-inhibiting hormone, Lancet, 2 (1974) 1106--1109. 2 Uvn/is-Wallensten,K., Efendic, S., Johansson, C., Sj6din, L. and Cranwell, P.D., Effect ofintraluminal pH on the release of somatostatin and gastrin into antral, bulbar and ileal pouches of conscious dogs, Acta Physiol. Scand., 110 (1980) 391-400. 3 Fiddian-Green, R.G., Farell, J., Havlicheck, D., Kothary, P. and Pittenger, G., A physiological role for luminal gastrin?, Surgery, 83 (1978) 663--668. 4 Uvn/is-Wallensten,K., Gastrin release and HCI secretion induced by electrical vagal stimulation in the cat, Acta Physiol. Scand., suppl. (1976) 438. 5 Schusdziarra, V., Harris, V., Conlon, J.M., Arimura, A. and Unger, R., Pancreatic and gastric somatostatin release in response to intragastric and intraduodenal nutrients and HCI in the dog, J. Clin. Invest., 68 (1978) 509-528. 6 Fiddian-Green, R.G., Pittenger, G. and Kothary, P., Effect of luminal somatostatin on acid secretion and gastrin release, Scand. J, Gastroenterol., 15 (1980) 305-309. 7 Skare, S., Hanssen, L.E., Hanssen, K.F. and Aadland, E., Somatostatin is present in human gastric juice; pentagastrin stimulates somatostatin content in plasma and gastricjuice in man, Regul. Peptides, suppl. 1 (1980) S106. 8 Ren6, E., Briskin, A., McArthur, K., Brazeau, P. and Isenberg, J.l., Intraluminal somatostatin in man. Effect of vagotomy, Gastroenterology, 82 (1982) 1160A. 9 Arnold, R., H/ilst, M.V., Neuhof, C.H., Schwarting, H., Becker, H.D. and Creutzfeldt, W., Antral gastrin-producing G-cells and somatostatin producing D-cells in different states of gastric acid secretion, Gut, 23 (1982) 285-291. 10 Accary, J.P., Vatier, J., Pointner, H. and Bonfils,S., Saline precipitation and chromatography technics applied to hog pancreas and gastric antrum; evidencefor immunoreactive gastrin, Horm. Metab. Res., Suppl. 5 (1974) 108-112. 11 Rorstad, O.P., Epelbaum, J., Brazeau, P. and Martin, J.B., Chromatographic and biological properties of immunoreactive somatostatin (IRS) in hypothalamic and extrahypothalamic brain regions of the rat, Endocrinology, 105 (1979) 1083-1092. 12 Chayvialle, J.A., Miyata, M., Reyford, P.L. and Thompson, J.C., Effects of test meals, intragastric nutrients and intraduodenal bile on plasma concentrations of immunoreactive somatostatin and vasoo active intestinal peptide in dogs, Gastroenterology, 79 (1980) 844-852. 13 Arimura, A., Groot, K., Taminato, T., Ertan, A., Akdamar, K., Ryan, J., Pienta, K. and Vinik, J., Somatostatin in human pancreatic and gastric juice, Peptides 2, Suppl. 2 (1981) 271-274. 14 Lehy, T., Peranzi, G. and Cristina, M.L., Correlative immunocytochemical and electron microscopic studies: identification of enteroglucagon, somatostatin and pancreatic polypeptide like containing cells in the human colon, Histochemistry, 71 (1981) 67-80. 15 Lehy, T., Voillemot, N., Dubrasquet, M. and Dufougeray, N., Gastrin cell hyperplasia in rats with chronic antral stimulation, Gastroenterology, 68 (1975) 71-82. 16 Dupas, J.L., Weiss, A.M., Capron, J.P., Brousse, N., Voillemot, N., Accary, J.P. and Potet, P., Evaluation quantitative, sur biopsies endoscopiques, des cellules fi gastrine dans la gastrite antrale, Gastroenterol. Clin. Biol., 4 (1980) 681--685.

143 17 Keuppens, F., Willems, G., Vansteenkiste, Y. and Woussen-Colle, M.C., Estimation of the antral and duodenal gastrin cell population removed by gastrectomy from patients with peptic ulcer, Surg. Gynecol. Obstet., 146 (1978) 400-406. 18 Takahashi, T., Shimazu, H., Yamagishi, X. and Tani, M., G-cell populations in resected stomachs from gastric and duodenal ulcer patients, Gastroenterology, 78 (1980) 498-504. 19 Arnold, R., Garbe, I., Koop, H., Mrnnikes, H. and Schwarting, H., Effect of antacid and H2-receptor blocker treatment on gastric endocrine cells, Scand. J. Gastroenterol., Suppl. I01, 19 (1984) 31-37~ 20 Delince, P., Willems, G. and de Graef, J., Antral gastrin cell proliferation after vagotomy in rats, Digestion, 18 (1978) 27-34. 21 Dunn, D.H., Decanini, C., Bonsack, M.E., Eisenberg, M.M. and Delaney, J.P., Gastrin cell populations after highly selective vagotomy in the dog, Am. J. Surg., 137 (1979) 111-115. 22 Lehy, T., Gr~s, L. and Ferreira de Castro, E., Quantitation ofgastrin and somatostatin cell populations in the antral mucosa of the rat. Comparative distribution and evolution through different life stages, Cell. Tissue Res., 198 (1979) 325-333. 23 Meikle, D., Taylor, K.B., Truelove, S.C. and Whitehead, R., Gastritis, duodenitis and circulating levels of gastrin in duodenal ulcer before and after vagotomy, Gut, 17 (1976) 719-728. 24 Horton, J.W., McClell, R. and Weger, R., Effect of parietal cell vagotomy on gastric emptying, Am. J. Surg., 143 (1982) 86-89. 25 Royston, C.M.S., Polak, J., Bloom, S.R., Cooke, W.M., Russel, R.C.G., Pearse, A.G.E., Spencer, J., Welbourn, R.B. and Baron, J:H., G cell population of the gastric antrum, plasma gastrin, and gastric acid secretion in patients with and without duodenal ulcer, Gut, 19 (1978) 689-698. 26 Piris, J. and Whitehead, R., Gastrin cells and fasting serum gastrin levels in duodenal ulcer patients, J. Clin. Pathol., 32 (1979) 171-178. 27 Uvn~is-Wallensten, K., Occurrence ofgastrin in gastric juice, in antral secretion and in antral perfusates of cats, Gastroenterology, 73 (1977)487-491. 28 Walsh, J.H., Richardson, C.T. and Fordtran, J.S., pH dependence of acid secretion and gastrin release in normal and ulcer subjects, J. Clin. Invest., 55 (1975) 462-468. 29 Jordan, P.H., A prospective study of parietal cell vagotomy and selective vagotomy-antrectomy for • treatment of duodenal ulcer, Ann. Surg., 183 (1976) 619-626. 30 Uvn/is-Moberg, K., Jhrhult, J. and Alino, S., Neurogenic control of release of gastrin and somatostatin, Scand. J. Gastroenterol., suppl. 89, 19 (1984) 131-136.